..

A.I.D. WORKING PAPER NO. 99

THE LOGICAL FRAMEWORK APPROACH TO

PROJECT DESIGN, REVIEW AND EVALUATION IN A.I.D.:

GENESIS, IMPACT, PROBLEMS, AND OPPORTUNITIES

Richard Ray Solem

April, 1987

center for Development Information & Evaluation

Agency for International Development
Washington, D.C. 20523
"

The Logical Framework Approach To

portfolio Design, Review and Evaluation In A.I.D.:
Genesis, Impact, Problems And Opportunities

Introduction

I. Genesis of the Logical Framework

II. Impact Of The Logframe

A. Impact Within A.I.D.
B. Im~act Among Ot.her Donors
C. capital Verses Technical Assistance Impacts.

III. Problems With Logical Framework Publication

A. Comments From Bureau Evaluation Officers
B. Analysis Of T;le Logical Framewor k
1. Project Structure Column
2. Important Assumptions Column
3. The Logical Framework To A Blueprint..

IV. Opportunities For Additional Logframe Applications

A. Capital Transfers
B. The CDSS Process
C. Agency-wide Policy Development
• Ii~TRODUCTION

A.I.D. installed the logical framework sy~tem in 1971 as a

response to a long-felt need for more effective evaluation of
substantive project impacts. Over the succeeding years, the
logical frameworl~ (logfrarne) methodology has had a profound
impact on A.I.D. 's project design system, and an important
though lesser impact on A.I.D.'S approach to project review and
) evaluation.

Today, fifteen years ~it~L the logframe was formerly adopted by

A.I.D., the system has been thoroughly institutionalized. It
provides ~ practical framework for project design i review and
evaluation that, used properly, can greatly increase the
likelihood of success in these tasks. Further evidence of the
logframe's usefulness is apparent in its adoption by most of
the major Western donor agencies in one form or another.
Impacts differ in all agencies, however, depending on the
emphasis given to technical verses capital assistance.

Despite this generally positive scenario, serious problems with

logframe applications continue to plague A.I.D. Depending upon
one's role in the organization (designer, reviewer or
evaluator) and upon one's portfolio (e.g. bilateral direct
investment, matching grants to PVO intermediaries or centrally
funded research), the usefulness of the logframe vaties
dramatically.
 At the same time, the opportunities for achieving great~r

effectivene~s in pursuit of A.I.D. 's mandate by applying the

logframe approach in new areas begs to be addressed.
Applicability of the logframe approach in design of capital
transfer projects, country development strategy statements,
even in Agency-wide policy-making, should be assessedw

In this ~aper the genesis, impact, problems ~nd opportunities

of the logframe approach are explored with a view toward
achieving greater appreciation of (1) what has transpired, and
(2) work still to be done if the logframe approach is to
realize full potential for facilitating the business of
•
economic development.
 I GENESIS OF THE LOGICAL FRAMEWORK

In reviewing the origins of the logical framework two

individuals stand out as key players; (1) Herb Turner, the
A.I.D. man who sensed the need for a better system for
conceptualizing and structuring projects and thus promoted the
study which led ~o the logframe, and (2) Leon Rosenberg, the
contract employee who served as principal author of the logical
framework in the course of reviewing A.I.D. 's project
-I evaluation system for Fry Consultants, Inc.

Herb Turner, a career A.I.D. man since Marshal Plan days,

traces the following key events as paving the way to
transforming A.I.D. 's approach to project design and evaluation:

The Kennedy Task Force's "Research, Evaluation and

Planning Assistance Report- of 1961, which led to the
creation of a special unit in A.I.D. for research into
economic development.

The wLincoln Report- of 1964, which increased

awareness of the need for internal evaluation in A.I.Da
 Appointment of Joel Bernstein, a long-time advocate of
project evaluation, to be head of the Technical
Assistance Bureau. Berstein was a constant advocate
of evaluation, and author of the 1968 ·Report to the
Administrator on Improving A.I.D. 's Program
Evaluation W in which A.I.D.'s system of (1) A.I.D./W
and Mission Evaluation Officers, (2) ·Spring Reviews·
of policy issues, and (3) the PROP, PIP and PAR system
were proposed.

These events are viewed as important, though they are not

related to the specific concept of the logical framework,
because they helped A.I.D. come to terms with the need for a
unified system for project evaluation, thus preparing a
receptive environment for the changes that were to occur.

In 1969, as a follow-up to Bernstein's report, specifically

with regard to the PAR installation, A.I.D.'s Office of
Evaluation contracted with Fry Associates, a D.C. consulting
firm, to review A.I.O.'s experience with its then new project
evaluation system. Fry's instructions were to look at problems
with A.I.D.'s recently installed Project Appraisal Report (PAR)
to determine why it didn't work better.
 In Fry's July, 1970 report to A.I.D., it was concluded that
A.I.D. 's problem with accomplishing an effective evaluation
system wasn't with the PAR so much as with the project design
process. To arrive at one's target destination, one needs to
plot a course that leads there. A.I.D. was failing to plot a
course that could lead to its target objectives. In this
report the essential first elements of the logical framework
dialectic, linking project design and evaluation, were laid out.

Turner, as the ninside man- who saw the need for improvement in
A.I.D. 's system, recognized the truth and simplicity of
Rosenberg's hypothesis, and paved the way for a follow-up
contract to d~velop the concept further. This follow-up
contract was with Practical Concepts, Inc. (PCI), to which
Fry's key logframe team moved.

The fruite of the first PCI report on thb logical framework

were delivered in 1970, and they created enough of a stir among

•
A.I.D. 's small cadre of project evaluators and evaluation
supporters to enable a decision to go forward with a
thoroughgoing overhaul of the Agency's approach to evaluation
along the lines of the logical framework dialectic. Another
contract was signed with PCI to design and carry-out a
large-scale training effort focussing on the principal
USA.I.D.s world-wide.

Consicering the magnitude of innovation the logframe approach

represented for A.I.D., the rapidity of acceptance was
remarkable. This qUick acceptance by A.I.D. employees was due
to a combination of top-level support, thorough planning and
follow-through by the staff of A.I.D.'s Office of Evaluation,
and assignment of unusually capable contract staff to carry-out
training in the field Missions.

A large scale field training/installation exercise was

undertaken in 1970, with the support of the A.I.D.
Administrator. During a six month period, teams composed of
A.I.D. and PCI employees were sent into USA.I.D. Missions
world-wide to carry-out one-week, highly intensive training
exercises. The focus of the field visits was on being
helpful. The USA.I.D.s, aluays suspicious of A.I.D./W
~nitiatives, were invited to select two of their on-going
projects for review from the logical framework perspective to
see how they might be better understood or redesigned. The
theory was that a helpful approach, dealing with the USA.I.D. 's
own problems, would m~ke the exercise ea~ier to relate to while
giving the training exercise a higher priority for Mission
staff.

The classic drill was Sunday arrival of the training team, a

Monday morning meeting with the Mission Director to discuss
plans, then formation of two Mission teams, each working with a
trainer, to analyze an on-going Mission project from a logframe
perspective. By Thursday the teams had done their separate
analyses and were united in plenary to compare notes. Wrap-up
was Friday, with the trainers en-route to their next
destination Saturday morning.

Following this six month long training initiative, A.I.D.

formally pronounced the logical framework approach the official
system for design and evaluation.

Subsequent to the field visits, classes for training in the

logical framework were made a part of the A.I.D./W program
management curriculum with a view toward training as many of
A.I.D.'s direct hire personnel as possible. These were
popular, and were soon opened to selected other-donor,
developing country, PVO and university representatives.
A.I.D. 's Program Development and Evaluation seminar was one
week in length and covered the entire program documentation
cycle from voting of the Foreign Assistance Act at the
beginning to writing of PIO/Ts and evaluation reports at the
 end. Actual time on the logical framework itself, therefore,
represented only a portion of that week's training - perhaps
two of the five days.

Training in the logical framework was also provided for all

incoming classes of International Development Interns. In this
~ instance, however, such training went well beyond the one-week
course to include substantial parts of their entire three- to
four-week preparation in project design.

An important final boost to acceptance of the logical framework

in A.I.D. came from the joining together of loan and grant
documentation in the early 1970~i. antil that point, there were
many A.I.D. staff, particularly representatives of the
development loan portfolio, who viewed the innovation with
skepticism. There had long been a division between loan and
grant professionals, and the loan and grant systems tended to
operate parallel to one another, with few intersections.
Whereas A.I.D.'S development grant documentation system had
minimal formal structure prior to 1970, the loan documentation
system was highly structured with checklists and special
requirements imposed by a variety of regulators. The logframe,
therefore, was initially seen by capital projects professionals
as just another troublesome checklist.
·As the basic documentation requirements were merged, however,
the logframe methodology entered the loan officer's mindset,
and soon proved to have a good effect on the loan development
process as well.
I

II IMPACT OF THE LOGFRAME

The impact of the logical framework within A.I.D., and the

international donor. community in general, has ceen profound.
In Part II thi~ phenomenon is traced through a close look at
A.I.D. itself, and then through a more cursory review of the
logframe's replication among other donors.

A. Impact Within A!D

The impact of the logical framework within A.I.D. has been

profound, particularly within the realm of project design and
evaluation. It has become so much a part of A.I.D.'s design
and evaluation process that the logframe voca~ulary is now
A.I.D.'~ vocabulary. One really cannot talk about design and
evaluation issues in A.I.D. without doing so in the terms of
the logframe's Goal-purpose-Output-Input dialectic.

At the same time that the log frame a~proach has changed the
language of development, in certain other respects. the use and
prestige of the logframe matrix as a Jormal tool for design and
evaluation appears to be eroding. It is less likely that the
logframe matrix per se will be the focus of conversation at a
project review; for example. This diminution of interest"in
the formal instrument has a way of rippling through the entire
design and review process. The effect is essentially to send
the message -Don't spend undue time on the logframe because it
won't receive serious attent~on at the approval meeting.-
Inevitably, with such a message being sent, the people involved
in the design process only turn to the logframe if they view it
as useful to them; designers as a tool to create a well
integrated project, and reviewers as a tool to qUickly
understand a proposal. In other words, the logframe is
increasingly used only because it is useful rather than becau8e
it is a requirement. This is good in most respects, but given.
the importance of formal requirements in a :arge, corporate
structure ~uch as A.I.D., it also is threatening to the
long-term survival of the logframe approach.

Following are summaries of interviewee comments on the logframe:

1. Before and After. The A.I.D. project design system prior

to introduction of the logical framework consisted
primarily of a description of project inputs and outputs,
without much attention to broader program goals and the
relationship of the project to them. As a result of
introduction of the logical framework, there is
considerably more analytical rigor applied to the design
task than was customary earlier.
2. As a Tool for Project Review. Project reviewers say that
the log frame is very useful in facilitating their task
because it gives them a short summary to turn to, presented
in a format with which they are familiar. When it is not
possible to read an entire paper, and it normally is not,
the log frame enables the reviewer to understand a proposal
with minimal time investment.

3. Logfrarne Training. Most of the design officers incerviewed

had received formal logframe training, largely through
A.I.D. 's IDI program. Those who had not seemed no less
enthusiastic about the tool. The typical comment was ~I

use the log frame because it is useful,' not because anyone

is asking for it.-
4. Effect of Portfolio Changes. Many of the people spoken to
indicated that the logframe is less used today because of
the growth of program loans documented with the PADF, which
instrument does not require any specific analytical
processes. Some shrugged their shoulders, suggesting that
if you're just trying to transfer funds one needn't bother
with the rigor demanded by log frame analysis. Others said
that the logframe system would help us to do a better job
of designing the program loans, particularly where
systematic policy reform is an expected result of the
assistance.
Virtually. all indicated, however, that people involved in the
review and approval process do not care much about the logframe
anymore. It isn't asked for at many review meetings, and is
even less frequently discussed~ The consensus seems to be that
unless A.I.D. personnel are taught the value of the tool, it
will fade in importance and use.

5. Broader Applicability. Several of those interviewed

expressed a need to use the logical framework tool to make
a linkage between the CDSSs and Mission por.tfolios. It was
felt that often coordination is lacking.

6. Importance of Missions. Most agreed that if there is to be

any effective reinvigoration of the logical framework as a
tool for design, review or evaluation, the action will have
to be in the field missions given that under current r~~es

over 90 percent of portfolio decisions are made there.

As indicated in 3 and 6 above, though the logframe has had a

profound and salutary effect on A.I.D.'S project design, review
and evaluation processes, that impact threatens to fade. While
still regarded as useful to those who work with it, and still a
formal project design requirement, the logframe matrix is less
called-for and focussed-on in the review process. Inevitably,
in a highly structured system such as A.I.D.'s, if the boss
stops paying attention to something, sooner or later the
employee will follow suit.

What is the cause of this diminution of interest in the logical

framework by A.I.D. managers? It seems to have two causes:

1. The large-scale switch from well defined projects to

capital transfers. These capital transfers have developed
their own system of documentation, emphasizing description
verses analysis, and there is no logframe requirement.

2. A six year hiatus of logframe training (from 1980 till

1986), leaving a generation of recently employed
professionals uninformed as to the power of the logical
framework approach to portfolio design, review and
evaluation.

B. Impact Among Other Donors

Other donor agencies were exposed to the logical framework soon

after A.I.D. began its own assimilation process. Tt!is exposure
happened spontaneously in the field as A.I.D. and other donor
colleagues collaborated on project design tasks. It also took
place in a formal, directed fashion as selected other-donor
professionals were invited to participate in the periodic
A.I.D./W logframe training programs.
 The reactions of other-donor personnel were generally positive
from the outset, with some eager to try it out, and others

,
electin~ to wait and see.
.~

In the former category is the German technical cooperation

agency, the GTZ. In 197_ the German government contracted
Practical Concepts, Inc., developer of the logframe, to install
such a system in their technical assistance program. Over the
course of a __ year contract, the GTZ carried out extensive
development and training efforts, devolVing a system in some
ways more sophisticated than A.I.D.'s own, the principal
difference being use of some corresponding -decision tree W
concepts.

Another early acceptor was the Canadian International

Development Agency (eIDA), however it was not long before the
logframe spread to most of the U. N. Development system, with
the United Kingdom Office of Development Assistance being the
last to employ the logframe in its programming.

The form that the log frame took, and how it is used, varies
somewhat from agency to agency. Moderate changes in the matrix
lables, and differences in emphasis (e.g. from use as a design,
implementation or evaluation tool), cause some appearance of
distinction. At bottom, however, the basic power of the
logframe remains intact in all applications; that is its
ability to show causality in the Project Structure Column, and
dependency on exogenous variables in the Assumptions Column.
 C. The Capital Assistance Hold-Out

Curiously, the significant hold-outs against acceptance of the

logical framework approach have come from that group of
economic development practitioners which theretofore had been
doing the best project design and monitorship work - the
capital projects people.

In A.I.D. the reluctance of the capital projects {lending}

people to accept the logframe was finally overcome through
merging of the loan and grant documentation requirements in
1975. A similar phenomenon occurred in Germany where the
technical assistance {GTC} and capital assistance {KfW}
agencies operate separately, however since both operate under a
common ministry (Economic Cooperation), it appears that the
capital assistance people will soon be prevailed upon.

The most notable hold-out, however, is the World Bank. Though

Bank personnel are generally familiar with the logframe
approach, and many use it on an informal basis, there is as yet
no serious effort to formally adopt the system.

How can one explain the general reluctance of capital projects

people to adopt the powerful and seemingly generic logframe
dialectic. The answer, it seems, lies not in the logframe
itself, but rather in the approaches to ascertaining project
viability for capital projects.
 Lender personnel, it seems, have a very well defined set of
practices· developed over decades of investment banking
J

practices. A good project is seen as one which generates

sufficient resources to repay the loan, and a little more to
ensure economic growth. The focus is on repayment, and the
economic growth test (generally given through some sort of
benefit/cost analysis) is viewed as a valid test of feasibility
by itself. If benefits exceed costs, the project can be
assumed to be good. The question which segment of the
recipient popUlation ultimately received the benefits tends to
be a secondary consideration in such analysis.

Contrarily, logframe practitioners, perhaps because of the

political climate in the American aid program at the time of
its inception, tend to be highly focussed first on .........
who
benefits, with the degree of benefit being of secondary
interest. Thus was born the division between the two camps,
and so it continues to some degree even in A.I.D., and to a far
greater extent in the World Bank.

The logical solution, of course, is a merger of the two

methodologies to the extent possible. This is what A.I.D. has
attempted since 1975, and though the marriage has benefited
both sides it has not been smooth for either. Classic
casualties include farm credit projects (on the capital
assistance side) rendered inoperable because of excessive
channeling of loans to borrowers generally unable to repay, or
because of administrative costs (to force such channeling) that
proved greater than the project could afford. An example of
the opposite distortion, (e.g. efforts to quantify impacts of,
say, a health project, into a benefit/cost analysis) have
likewise somewhat discredited valid economic analysis
techniques through misuse.

On reflection, what A.I.D. and its other donor colleagues are

witnessing in the battle between capital and technical
assistance personnel on the logframe arena is not unlike the
conflict between policy and project personnel in general. The
logframe has provided a common field on which players from the
extremes of the economic development process can meet to
discuss their common purpose. Given the variety of players in
the process, it is perhaps fitting that the dialogue should be
at times difficult.
 III PROBLEMS WITH LOGICAL FRAMEWORK APPLICATION

In Part III we look first at general comments from Evaluation

Officers from each of A.I.D.'S five project implementing
bureaus, assessing such comments for points of agreement.
Following this, direct analysis of the logframe is undertaken,
looking at it on a component by component basis. This
analysis, in turn, is wrapped-up with summary conclusions with
respect to the overall log frame concept - the logframe as a
blueprint for project design.

A. Comments From Bureau Evaluation Officers

In order to better assess the continued usefulness of the

logical framework in A.I.D., additional interviews were made
with ~n emphasis on application problems. Interviewees were
the Evaluation Officers from each of the five A.I.D. Bureaus
having program management responsibility; LAC, AFR, ANE, S&T
and FVA. Following is a summary of these interviews, and an
assessment of the implications for continued successful
logfrarne application within A.I.D.

(1) Latin America & Caribbean

-Limitations in effectiveness of the Logical Framework arise

more from improper use than from shortcomings with the
instrument itse1f.-
-LAC's current passion is Management by Objectives (MBO). In
the MBO framework, a logframe Goal is synonimous with an MB~

Objective. The MBO exercise in LAC is focussed on the Action

Plan: a step above projects and a step below the CDSS. The MBO
approach involves Indicators, Time Frame and Assumptions, as
the logframe deals with Indicators, Means of Verification and
Assumpti~ns.·

-Though the logframe is a good design tool, many LAC design

officers miss the mark by treating it only as -another annex·
to be dealt with. As a review tool it is effective when
properly prepared - gives a good grasp quickly, like an
Executive Summary. For evaluation it is also effective if
properly prepared in the first place.

(2) Asia & Near East

-ANE's major problem with the logical framework is inability to

apply it correctly. More training is needed. The training
program given by Management Systems International (MSI) is
excellent and should be funded more heavily.-

-Tendency is to use only the left hand column of the Logframe.-

WA good Logframe provides guidance to project implementers.
This suggests that it needs to be continually rewritten, as
circumstances change. Logframing is really a management by
objectives exercise w •
(3) Africa

WThe great strength of the Logframe is that it helps people to

think through the Input, Output, Purpose and Goal linkages. w

WThe great disappointment of the Logframe is that it is so

often improperly used. People tend to dwell on how to fill in
the boxes rather than on the linkages themselves.-

-Logframes are also often improperly viewed as wblueprints W -

something that once made cannot be changed. This is a serious'
misuse of the instrument that gives an unfairly bad
reputation. Logframes should evolve even as project
circumstances evolve. Logframe revisions should be tied to
adjustments in a Mission's Action Plan. The CDSS provides
Mission Goals, the Action Plan describes Objectives
contributing to CDSS Goals, and project logframes are linked to
the same Goals and Objectives. w

WAn interesting approach to training might be a course

reviewing what A.I.D. has learned from use of the Logframe over
the past 15 years - how we can do better in light of this
experience. Case studies would provide a good teaching aid.-
 (4) Science & Technology

-The logframe is an excellent tool which is well adapted to

A.I.D. 's bilateral assistance program. In the case of S&T's
portfolio of mostly research projects, however, it is harder to
apply. We tend to supply Outputs to many Mission efforts, with
little clear linkage to specific Purposes and Goals. We also
have trouble using the Assumptions column.-

-For effective logframing of S&T projects the logframe may need

some special treatment. As a minimum, special training is
needed.-

(5) Food & Voluntary Assistance

-The logframe approach to project design is important and

useful, but we have had difficulty applying it to our portfolio
in FVA for some very good reasons:

PL-480 initiatives are approved by a government-wide

committee (DCC) which has no tradition of thinking in
project terms. They are influenced by (a) a desire to move
the commodities and (b) stated humanitarian objectives of
the legislation.
 Grants to PVOs don't lend themselves·easily to logframing
either because the whole idea cf "matching grants" is that
A.I.D. simply supports the work of institutions already
proven worthy. Such institutions tend to work in m~ny

countries on many projects, and A.I.D. support is general

in nature.
Despite these inherrent difficulties of application in FVA, the
Logframe is viewed as an important instrument of potential
utility. The trick is to figure out how to apply it."

"Recently FVA developed guidance with a view toward better

projectization of PL-480 activities. This should facilitate
logframing somewhat."

•
"With respect to matching grants to PVOs, we've got them using
"operating plans" now, which is a step in the right direction."

"The key to applying the Logframe to the FVA portfolio is

training - perhaps a special course focussing just on PL-480
and "matching grant" programs."

---------------

Assessing these and other interviews from the perspective of

problems with continued usefUlness of the logical framework in
A.I.D., the follOWing points can be made:
(1) The logframe tool itself is good; problems with it are
from misapplication by its users.

(2) Parts of the A.I.D. project por~folio lend themselves

less readily to logframe application than the classic
bi-lateral projects (e.g. S&T research and FVA
matching grants and PL-480 programs). The log frame is
also little used with bi-lateral capital transfers,
not because this type of activity presents application
problems so much as that the Agency simply hasn't
required logframe use for such activities.

(3) Agency, Bureau and Mission policy-making and

programming exercises may also be carried-out more
effectively if the logframe tool is employed to better
ensure awareness of interconnectedness between
projects, programs and policies.

(4) Additional logfrarne research and training are needed

to help Agency design, program and policy-making
personnel learn how to apply the logframe to their
activities more effectively.
 B. Analysis of the Logical Framework Instrument

Looking, now, at the logical framework instrument itself,

following is a brief assessment of perceived problems, element
by element:

1. Project structure Column

The genius of the logframe is its ability ~o help ordinary

people not given to seeing far beyond their immediate actions a
tool for projecting consequences well into the future.
Employed properly, it enables its users to function like
chessmasters, knowing with some degree of confidence the 2nd,
3rd and· 4th generation consequences of each move. The

-.!
input/output/purpose/goal statement of the logframe is really
nothing more than a formula for projecting consequences of
one's actions.
On a conceptual level this dialictic has not been too difficult
for A.I.D. personnel to digest. However, problems of
application have been considerable.

The first problem comes from argument over where to start.

The natural inclination of ·hands-on· project designers and

managers is to start from the bottom and work up, concerning
themselves with input/output/purpose linkages in that order or
priority. If the project is rice culture, for example, the
project manager is going to concern himself primarily with
bringing together the appropriate measures of land, labor and
capital (read farm land, farmers and a
seed/equipment/technology package) in timely fashion so to
produce a bountiful harvest.

The natural inclination of the policy-makers, on the e,ther

hand, is to start from the top and work down, focussing on
goal/purpose/output linkages in that order of priority. The
debate among policy-makers has an entirely different flavor,
therefore, even where the same project is concerned. More than
often it would treat issues such as equity and economic
opportunity vis a vis higher farm income, paying only the
•
slightest heed to rice culture itself.

This dichotomy of interests in the very same project highlights

both the genius and failure of the logframe approach to project
design, review and evaluation. The logframe is useful because
it provides a common playing field, and language, for the
players on both extremes of the overall economic develg~ment

team (project managers and policy-makers), yet it generally

fails to ensure a successful outcome. In the vast majority of
circumstances project designs that appeared to have adequately
addressed the goal/purpose/output/input linkages, or if you
prefer, the input/output/purpose/goal linkages, have come up
short. Some of the linkages are successfully made, but not
all. If you compare this shortfall with failing to complete
the entire span in a bridge, you get an idea of the
consequences of partial success. If the goal is only
accomplished with a complete crossing of the river, and your
project takes you 75% of the way, you may have constructed a
'very impressive span, as far as you went, yet done no good at
all with regard to goal achievemento

Where doe:-' the breakdown occu r, and why? This question takes
us to another part of the logframe instrument - the column
wherein the behavior of non-A.l.D. players is predicted.

2. Important Assumptions Column

The problem with systems for predicting outcomes is that they

only work where all variables are under control. If Player B
in a chess game, for example, could be re.Lied upon to respond
to everyone of Playe.r A'S moves in a. preCleterined way, it
would be relatively si.mple to develoI> a s~'stem by which Player
A could win every time~ Unfortunately, Player B cannot be
relied upon in that way. His counter-moves might be influenced
by factors as diverse as length of the game, how well he slept
the night before, even how he feels about Player A personally.
If it is difficult to predict behavior in one player in a game
as well defined as chess, imagine the difficulty of predicting
the behavior of large numbers of people and institutions in a
game as complex and changing as life itself. Is it any wonder
that A.I.D. practitioners have problems with the logframe's
Important Assumptions Column?

The recommended methodology at the outset, upon preparation of

the logframe by a project design team and its reviewers on the
policy-making side, is valid in practice and works reasonably
well. At each level of the project structure
(inputs/outputs/purpose/goal) the assumptions critical to
realizing predicted linkages are defined. On one end the
assumptions might be (1) that the land, labor and capital for
successful rice production can be assembled in timely fashion,
and (2) that nature will provide wind, rain and sunshine at the
appropriate times.

Note: Inputs is the easy part, and already we are dependent

upon a variable totally outside A.I.D.'s control - the whims of
nature.

On the other end some enormously speculativ~ assumptions are

made; e.g. that rice buyers will appear on the scene in timely
fashion, that rice prices will be sufficient to ensure profit
to farmers, that the farmers benefiting will be those intended
to benefit, etc. In between the input and goal assumptions are
many more equally heroic predictions.
What do these problems with the Important Assumptions Column
tell us? . Does our failure here mean that we should abandon
efforts to define such variables altogether and give up on
efforts to predict input/output/purpose/goal linkages?
Alternatively, perhaps it means that A.I.D. should only operate
in controlled environments where there are no exogenous
variables - much like a scientific laboratory~

The answer, it seems, is a compromise. To be effective,

development agencies must adapt themselves to the environments
in which they operate. At the same time, complete abandonment
of scientific method would be unwise. Just as a successful
chess player must know many strategies, and constantly adjust
between them depending upon how his opponent reacts (chcanges
the game's assumptions), so too must A.I.D.ls development
practitioners constantly adapt to the changing environment in
which the A.I.D. p[o~ect is being played out.

In short, A.I.D. must continually review the assumptions upon

\lrhich the input/output/purpose/goal linkages are depend~Lnt, and
adjust its game plan accordingly. Logical? Yes: Common
practice? No:
3. The Logical Framework As A Blueprj~

The problem A.I.D. practitioners are up against here is that

project designs, and the logical frameworks through which they
are framed, are treated like blueprints which cannot be
changed. This, of course, makes no more sense in an economic
development context than it does in more traditional
applications for blueprints. Let's look at this misuse of the
logframe blueprint from the perspective of an architect.
Imagine, if you will, a home built from a blueprint providing
utility and structural integrity for five rooms on one level •.

Now let's assume that the circumstances surrounding the horne's

occupants change, and they elect to, say, add a floor above and
a garage at one end. Since the occupants don't want to be
bothered with the hassle of reviewing the basic house plans,
they just begin to add-on, in the process changing load factors·
for electricity, hot water, heat and air conditioning, weight
on frame, outside drainage, etc.

Is there anything wrong with adapting the home to changing

circumstances? No= Are the changes likely to cause problems
if made without making corresponding adjustments in basic
design features? Yes!

-35-
This example illustrates the problem frequently encountered
with the logical framework. Project circumstances change, so
on a day-to-day basis adjustments are made. Rarely, however,
are these adjustments assessed in light of the basic project
blueprint, the logframe.

The why of this curious behavior also seems to bear some

resemblance to the illustration of architectural and
engineering plans for a home. To formally change a homefs
design specifications there is a necessity of involving certain
public officials - electrical, water and fire inspectors, to
name a few. This is often viewed as an unnecessary bother for
just one more piecemeal adjustment.

The perspective of the A.I.D. project manager is similar. With

the day-to-day circumstances surrounding the project in
constant turmoil, the manager is tempted to view basic design
reassessment as a time-consuming distraction involving folks
from the Program Office, and the Directorfs Office, who really
are not sufficiently -in touch- to be of genuine assistance.
So the project manager innovates by instinct, Wflying by the
seat of his pants.-

What is the remedy?

-36-
The first step is to communicate clearly to all logical
framework users some obvious truths:

Is the logical framework like a blueprint? Yes:

Can/should it be changed as project circumstances change?

Yes:

If these two messages are clearly understood, then execution of

the policy is simply a matter of mechanics. Needed is a way to
encourage project managers to review the blueprints on a
continuing basis.

Indicated, it would seem, is to build logframe review into the

regular reporting process - those monthly or quarterly reviews
by which managers keep their Mission Directors informed of
important events. Such a review would give greatest emphasis
to the Important Assumptions Column, because it is there that
changing circumstances have their most direct impact. It
WOUld, however, look likewise to the Input/Output/Purpose/Goal
linkages, to Achievement Indicators, and to Means of
Verification.

-37-
•

Does this sound like a heavy burden to put on A.I.D.'s hard

working project managers? Yes= Is it reasonable? Yes= Even
in the home rehabilitation business it is unpleasant, and often
intimidating, to have to review the compliance of one's
residence with basic building code requirements, but only in so
doing can one avoid large-scale non-compliance as one
adjustment is added to another over the years.

-38-
 IV OPPORTUNITIES

In Part! we trac~d the birth of the logframe and its initial

application within A.I.D. In Part II we discussed the
logframe's impact on A.I.D., and on the international donor
community in general, affecting both the form of project
documentation and the language of the discourse. Then in Part
III, having acknowledged these successes, we proceeded to
reveal grave inadequacies whic~, left uncorrected, threaten to
discredit the logframe approach altogether.

Now, in Part IV, we will focus once again on what the logframe
~~, leaving aside a lot of real world application
shortcomings. With this positive perspective, we will explore
other opportunities within A.I.D. wherein thoughtful
application of the logical framework methodology might lead to
better results.
A. Cash Transfers

Based upon trends in funds obligated, the Whottest W segment of

A.I.D. 's overall portfolio is that characterized by nonproject
assistance, i.e. cash transfers of one form of another. This
category of project has grown with the increasing importance of
A.I.D.'s -Economic Support Fund (ESF) portfolio - funds tied
closely to U.S. political interests abroad. In FY 1986 some 68
percent ($4.91 billion) of the A.I.D. portfolio
overall was allocated to SSA with 71 percent of that amount
($3.49 billion) being obligated though cash transfer mechanisms
- roughly 49 percent of A.I.D.'s total $7.18 billion bUdget.

A critical difference between nonproject assistance and

A.I.D.'s project assistance portfolio is the required
methodology for documentation. Conventional A.I.D. projects
(largely the Development Assistance portfolio) are prepared per
Handbook 3 guidance for preparation of the Project Paper (PP),
the core of which is the logical framework. Nonproject
assistance (largely the ESF portfolio), on the other hand, is
governed by Handbook 4 guidance for preparation of the Program
Assistance Authorization Document (PAAD).

PAAD guidance in Handbook 4 does not lack detail. Indeed, it

is in many respects more detailed than Handbook 3 guidance for
preparation of project Papers. Direct information requirements
for, for example, economic policy analysis, pipeline analysis,
balance of payments analysis and financial analysis provide the
writer with opportunities to present volumes of -learned-
information.

Such presentations may, indeed, deal with the proper issues and
thus assist sound investment decisions. It is often difficult
to tell, however, what with all the information. Missing from

-40-
the PAAD presentation is a requirement for rigorous analysis -
for a clear demonstration of the relationship between what is
proposed and overall Agency and Host Country goals, and
likewise for identification of the many exogenous variables
(assumptions) that could render the proposed nonproject
assistance ineffective.

It is quite possible with PAAD analysis to Wwrite around w the

difficult issues, merely presenting descriptive material in
response to wchecklist type W questions. The result is often
impressive appearing documents that (1) are hard for reviewers
to Wget a handle on w and (2) fail to deal with the issue of
genuine usefulness of the proposed assistance.

It is possible, in a word, to wwrite around w the difficult

issues, and to focus on that part of the analysis that is
convenient to treat -~ the sort of selective focussing that was
rendered impossible through imposition of the logical framework
on project assistance documentation.

There is a persuasive school of ,hought, however, that argues:

·So what? If it has been determined by Congress to provide
funds for political purposes, why should A.I.D. impede the
process with complicated analysis and documentation
requirements? Why not just send the money and use staff time
on jobs where they can better influence the outcome?W

-41-
The answer, it seems, lies in a compromise. There are ESF
capital transfers that are genuinely, 100 percent
pre-determined, from which it is not appropriate to attempt to
squeeze additional influence. Examples would be -base rights·
concessions where the capital transfer is really nothing more
than rent. These, however, are the minority of such projects.

In many cases the capital transfer is tied to certain

adjustments in economic policy; program loans, for example.
Where this is the case, analytical rigor in document
preparation is absolutely essential. There should be no
·writing around· the issues; no SUbstitution of descriptive
material for analytical rigor. In these instances, the
logframe is ideally designed to force recognition of the causal
linkages (Project Structure Column), on one hand, and critical
exogenous variables (Important Assumptions Column) on the other.

Another potentially useful application of logframe techniques

to nonproject assistance (capital transfers), is where there
are provisions for generation of local currency to be jointly
programmed - e.g., many commodity import projects. In such
instances, establishment of project selection, analysis and
documentation standards based u~un use of logframe techniques
could have very positive effect by (1) improving the quality of
secondary projects funded and (2) training host country
personnel in use of the logframe.

-42-
 B. The CDSS Process
-
Programming in A.I.D. takes place on several levels. At the
base level we have projects, each of which is a discrete, often
highly complex development program. Up till this point, our
discussion of log frame techniques has dealt entirely at this
level.

At the next level, missions prepare an Annual BUdget Submission

(ASS), which serves as a memorandum to the U.S. Congress
outlining what is being done in each Mission and Bureau and
what is proposed. The format for the ASS is largely
descriptive, and advocative in approach. The ASS process is
not viewed by A.I.D. as a process for deciding what should be
done, but rather as a process for communicating to the Congress
what A.I.D. wishes to do. It is fitting, therefore, that the
format for th~ ASS should be determined largely by what
Congress is looking for, and that the style is advocative.

At the next level, however, there is a programming process in

which analytical rigor is of critical importance -- that is the
CDSS process.

Each year a handful of A.I.D. missions are asked to prepare an

overall review of their portfolios, current and planned
projects alike, and review them in A.I.D./W with Bureau and
Agency l~adership. This exercise is intended as an opportunity
for Missions to step back and take a thoughtful look at what
they are doing, and for A.I.O./W's policy people to have an
opportunity to aid in the process by sharing insights from
their broader, regional and world-wide perspectives on
development. The COSS process is potentially a very valuable
planning exercise.

The reality of the CDSS process, however, is different.

Despite concerted efforts by A.I.D./W management staff to
provide guidelines that will result in genuinely analytical
CDSS submissions, the results are often disappointing.

Too frequently the Missions respond to CDSS guidelines with

documents describing each sector of the host country economy
and providing -bullet-like- narratives of what is being done,
or will be done, to address the problems. What is generally
lacking is the argumentation to show linkage between the
problems posed and solutions offered.

A.ItD./W managers have attempted to resolve this by writing

detaile~ CDSS guidance cables asking Missions to provide an
analysis of the sector problems, and a discussion of why one
approach is favored over another. Although this instruction is
clearly stated in paragraph 4 of the 1987 CDSS guidance cable,
the paragraph tends to be lost among a variety of additional,
more specific instructions.

-44-
Needed, it would seem, is a proposed analytical framework that
cannot be ignored because it itself is an outline for the
analysis. The ogical framework matrix seems ideally suited for
structure of CDSS presentations. Because it is all inclusive,
one cannot ·write around· the logframe. Employed as a CDSS
structure it would force analysis of the causal relationships
between problems identified and projects proposed, and it would
force identification of critical exogenous variables. Because
A.I.D./W and mission staff are already familiar with the
logframe in the project design context, it would simplify the
CDSS review and discussion process by providing a common frame
of reference, both in the Missions and in A.I.D./W. Less time.
would be spent on discussing ·what A.I.D./W wants· and how to
respond fully to the CDSS gUidelines.

C. Agency-wide Policies

The final A.I.D. activity in which potential for application of

logframe techniques is discussed is Agency-wide policymaking.
By policy-making is meant that general process of discussing
and resolving future directions and emphases for the Agency.
An important recent example of policy-making is the evolving
emphasis on ·privatization· in A.I.D. 's development
initiatives. Another even broader example would be evaluation
of the Agency's -four pillars· emphasis.

-45-
If policy-making is, indeed, such a broad term in A.I.D., how
can one organize to apply logframe techniques .to it? Indeed,
is it even a good idea? Since policy-making in A.I.D. involves
close coordination with Congress, and the State Department,
both of whom are often in a position to dominate the dialogue,
is it realistic that logframe rigor can be applied to the
process.

These are valid reservations, of course. It is difficult to

define the terms for a dialogue with one's boss. Yet perhaps
in this very problem lies the rationale for attempting to du so.

No one in the A.I.D. policy-making process opposes use of

•
logic. Given this, why are there disagreements about what
policies should be? The answer is that different parties to
the dialogue often have different base assumptions about
priorities, host country context, and the like. These
differences may never come to light because they go unstated.
The potential usefulness of the logframe approach to this
policy dialogue is that it urges analysis of all critical
linkages and assumptions.

Is it realistic to think that A.I.D. could get its counterparts

to deal with development policy issues in such an analytical
context? This can probably only be determined over time.
Needed first, however, is for A.I.D. 's policy-making personnel

-46-
to take such a focus themselves. Has anyone ever analyzed the
concept o~ privatization in thp. logframe format, attempting to
reconcile it with stated A.I.D. objectives? Would this be a
useful exercise in helping A.I.D. personnel to understand the
policy, and how it might be applied? Would our colleagues in
the department of state benefit from, say, a workshop built
around analyzing this policy in a logframe context?

The question of log frame applications to general policy-making

is difficult, indeed. If serious thought is to be given to
exploring this possibility, an experimental workshop might be
a good first step.

-47-
•

CONCLUSIONS

The logical framework approach in A.I.D. has been very

successful, so far as it has been utilized. Quality of
A.I.D.'s Development Assistance project designs has improved
markedly, and the review and evaluation processes ~ave been
facilitated as a result of the better designs.

For the logframe to -hold its own,- however, continued logframe

training for A.:.D. employees is a must. The hiatus in such
training from 1980 till 1986 has taken a toll by introducing an
entire generation of middle and senior managers to A.I.D.
without logframe familiarization. Such employees, often in
positions to participate in project review meetings, have shown
little interest in the logframe approach, and thus have lowered
the importance of log frame analysis in the design and review
process.

-48-
For the logframe to realize its potential for serving A.I.D.'s
changing portfolio requirements and policies, additional
investments should be made in two areas:

(1) A contract, and perhaps several workshops, to study

the potential effectiveness of logframe techniques to
(a) aid in analysis and design of certain capital
transfer projects currently exempted from Handbook 3
documentation guidelines, (b) preparation and review
of CDSS submissions and (c) development and review of
broad new Agency policy initiatives.

(2) Revision of the logical framework training module to

accomodate broader applications to such new activities.

-49-
 A critical difference between so-called capital transfer
projects und A.I.D.'s project assistance portfolio is the
required methodology for project documentation. Conventional
A.I.D. projects (largely the Development Assistance portfolio)
are prepared per Handbook 3 guidance, the core of that guidance
. being the logical framework. Nonproject assistance (largely
the Security and supporting Assistance portfolio), on the other
hand, is governed by Handbook 4 gUidance for preparation of the
Program Assistance Authorization Document (PAAD).

PAAD guidance in Handbook 4 does not lack detail. Indeed, it

is in many respects more detailed than Handbook 3 guidance for
preparation of Project Papers. Direct information requirements
for, for example, economic policy analysis, pipeline analysis,
balance of payments analysis and financial analysis provide the
writer with opportunities to present volumes of -learned-
information.

Such presentations may, indeed, deal with the proper issues and
thus assist sound investment decisions. It is often difficult
to tell, however, what with all the information. Missing from
the PAAD presentation is a requirement for rigorous analysis -
for a clear demonstration of the relationship between what is
proposed and overall Agency and Host Country goals, and
likewise for identification of the many exogenous variables
(assumptions) that could render the proposed nonproject
assistance ineffective.
It is quite possible with PAAD analysis to ·write around- the
difficult issues, merely writing descriptive material in
response to ·checklist type- questions. The result is often
impressive appearing documents that (1) are hard for reviewers
to -get a handle on- and (2) that fail to deal with the issue
of genuine usefulness of the proposed assistance.

-51-